Trajectory Planning Method for Driverless Sweeper in Unstructured Environment

In the paper, we propose an approach for full coverage path planning and obstacle avoidance trajectory planning in the unstructured environment, applied to the driverless sweeper with Ackermann-like steering. The approach uses the boustrophedon cell decomposition method to solve the full-coverage path planning problem while using DWA (Dynamic Window Approach) and A* and a finite state machine to solve the planning problem of obstacle avoidance trajectories. A kinematics model of the Ackermann-like steering structure sweeper is established to make the obstacle avoidance trajectory satisfy the kinematic constraints. The method designs a corresponding trajectory evaluation function to make the DWA applicable to the sweeping scenario. And a finite state machine model is designed for the decision problem in obstacle avoidance. Finally, using a physical sweeper of 1.5m x 1.0m x 1.2m size in a factory environment verifies the effectiveness of the proposed method.